Google AI Mode traffic hides inside standard search data by default. Here's what Google has actually rolled out to separate it, and what still doesn't work.
Tracking visibility in Google AI Mode specifically is harder than it should be, because for most sites, AI Mode and AI Overviews impressions are still blended into standard Web search results in Search Console, with no separate filter. Google announced dedicated generative AI performance reports on June 3, 2026, showing impressions from AI Overviews and AI Mode broken out by page, country, device, and date, but this rolled out to a limited subset of sites for testing, not universally, and as of August 2026 many site owners still report no separate AI Mode filter available in their own Search Console property, with AI-feature traffic still folded into the standard Web search type.
For sites that do have access, the new report is genuinely useful but incomplete. It shows impressions only, how often a site's URLs appeared inside AI Overviews, AI Mode, or generative AI features in Discover, broken down by page, country, device, and date. Clicks and CTR are not included in the initial rollout, which means it answers "were we shown" but not yet "did anyone actually click through," a meaningful gap for anyone trying to justify budget with a hard number.
For sites without access to the dedicated report, the practical workaround is inference rather than direct measurement: watching for the signature pattern of impressions holding steady or rising while click-through rate falls on informational queries, which is the pattern an AI Overview or AI Mode result typically leaves behind in standard Search Console data, since it satisfies the searcher's question without requiring a click through to the site. This isn't a precise measurement, it's a reasonable approximation, and should be treated as one.
Combining this with the GA4 approach covered in how to check and track whether AI systems know your brand gives a more complete, if still imperfect, picture: GA4 catches referral clicks from chat-based tools like ChatGPT and Perplexity, while the CTR-drop pattern in Search Console is currently the best available signal for Google's own AI Mode and AI Overviews specifically, at least until broader access to the dedicated report arrives.
Google has published data indicating AI Mode queries average roughly three times the length of a traditional search, and answers are frequently built from multiple sources at once through what Google calls query fan-out, meaning a single AI Mode answer can draw on and credit several different URLs in one response. This makes simple position tracking a poor proxy for AI Mode visibility specifically, a page can contribute to an AI Mode answer without ever registering a traditional top-ten ranking for the query that triggered it.
This is part of AI Search & SEO. This connects directly to how to check and track whether AI systems know your brand, which covers the GA4 side of this same measurement problem.
Trying to make sense of your own AI visibility data? Get in touch.